Job Market Paper:
Governments often mandate that private firms provide unprofitable services. Such mandates can have large benefits for the targeted population, but by increasing provider costs they may reduce overall supply, with adverse impacts on others. In this paper, I study the impact of a mandate requiring commercial banks in India to open 25% of their new branches in unbanked rural villages on the size, geographic distribution, and profitability of the national branch network.
Using novel, comprehensive administrative records of branch licenses and municipality-level Census and geospatial data, I first document the scope of the post-2011 rural branch expansion. In the five years after the reform, banks opened over 11,000 branches in unbanked villages, home to about 6% of the total unbanked population. The rural branch expansion was deep as well as broad. The median village banked after the reform has a 40% higher poverty rate than rural villages first entered in the five years prior.
I then use an economic model of branch entry to study the causal effects of the reform. I estimate banks’ profits, compute their regulatory compliance costs, and simulate equilibrium entry probabilities and profits under counterfactual policies. I find that, compared to a free-entry counterfactual, the mandate shifts entry from banked to unbanked markets roughly one-for-one, with disproportionate losses in smaller banked markets. Bank profits from new branches fall by 26%. Returning to the 67% unbanked share enforced in the 1980s would produce 50% more rural entry, with catastrophic impacts on banked entry and profitability.
Augmenting the baseline mandate tradable permits makes the mandate 11% less costly for banks but does not result in additional branches. Private banks become less constrained and replace public sector branches less than one-for-one, resulting in higher private sector profits and increased geographic concentration of bank branch networks.
This paper studies the implications of low energy prices today for energy efficiency and climate policy in the future. If adjustment costs mediate manufacturing plants' responses to increases in energy prices, incumbents may be limited in their ability to reoptimize energy-inefficient production technologies chosen based on past market incentives. Using U.S. Census microdata and plausibly exogenous variation in state energy prices, we show that the electricity prices manufacturing plants pay in their first year of operations are important determinants of long-run energy intensity. Plants that open when the prices of electricity and fossil fuel inputs into electricity are low consume more energy throughout their lifetimes, regardless of current electricity prices. We measure the relative contributions of initial energy productivity and capital adjustment frictions to the overall lock-in effect by estimating a model of plant input choice. We find that observed lock-in is mostly explained by persistent differences in the relative productivity of energy inputs. We discuss how these long-run effects of low entry-year energy prices increase the emissions costs of delayed action on carbon policy.
Work in Progress:
The Use and Misuse of Average and Marginal Energy Prices: Implications for Climate Policy, with Katherine Wagner
The goal of a carbon tax is to equate the marginal cost of reducing carbon emissions with the marginal social benefit of avoiding these emissions. However, the responses of firms and individuals to these taxes are often calculated using average costs commonly reported in economic data, rather than marginal costs. This paper quantifies the wedge between marginal and average electricity prices for industrial consumers and studies the implications for climate policy. First, we document that the marginal price of electricity paid by manufacturing plants is lower than the average price by constructing a panel of both average and marginal electricity prices using plant-level microdata from the U.S. Census and utility-level electricity rate schedules for over two hundred utilities. Second, we provide guidance on when average costs are an appropriate proxy for marginal costs by identifying economic and geographic characteristics that predict variation in the wedge between marginal and average electricity prices. Overall, we find that the standard use of average energy prices to calculate responses to carbon taxes may underestimate the energy price increases needed to meet emissions targets.